calculation for oxygen molecular

[matilda]

dear all,
I try to calculate the binding energy of molecular oxygen with vasp.4.6 in the GGA (pw91) with the PAW-GGA potential tooken in the vasp library. my calculations leeds to a value of 6.68 eV largely overestimated with the experimental value of 5.2 eV.
my incar for oxygen atom is :
PREC = High
ISYM = 0 ! no symmetry
ISPIN = 2 ! allow for spin polarisation
VOSKOWN = 1 ! this is important, in particular for GGA
ISMEAR = 0 ! Gaussian smearing, otherwise negative occupancies
SIGMA = 0.1 ! intermid. smearing width
AMIX = 0.2 ! mixing set manually
BMIX = 0.0001
NELM = 20 ! 20 electronic steps
ICHARG = 1
IALGO=48
PREC = High
ENCUT = 400

for the oxygen dimmer the incar file is :

SYSTEM = O atom in a box
ISTART = 2
ISMEAR = 1 ! Gaussian smearing
SIGMA = 0.1
ENCUT = 400
ISPIN = 1
ISYM = 0
VOSKOWN = 1
IALGO=48
ISIF =2
IBRION = 1
NSW = 10
PREC = High


can any one help me to resolve this.

Thank for your time and your help.

[admin]

1) for both, the molecule and the free atom, SIGMA is much too large, both, O and O2 have discrete levels in their electronic ground state configuration.
2) please also use Gaussian smearing for O2.
3) the ground state of O2 is the triplet state (ISPIN=2)
4) if you have not done so, please use the hard PPs

[rpnabar]

I was facing almost the same problem as matilda. A few questions:
1. I was using an ultrasoft GGA PP. How do I switch to the hard pp's.?

Thanks!

-Rahul

[bandy]

Hi, Hard PPs should be in the directory /element_h/POTCAR.Z.